-
Notifications
You must be signed in to change notification settings - Fork 721
Insights: iree-org/iree
Overview
Could not load contribution data
Please try again later
7 Releases published by 1 person
-
iree-3.6.0rc20250627 iree candidate iree-3.6.0rc20250627
published
Jun 27, 2025 -
iree-3.6.0rc20250628 iree candidate iree-3.6.0rc20250628
published
Jun 28, 2025 -
iree-3.6.0rc20250629 iree candidate iree-3.6.0rc20250629
published
Jun 29, 2025 -
iree-3.6.0rc20250630 iree candidate iree-3.6.0rc20250630
published
Jun 30, 2025 -
iree-3.6.0rc20250701 iree candidate iree-3.6.0rc20250701
published
Jul 1, 2025 -
iree-3.6.0rc20250702 iree candidate iree-3.6.0rc20250702
published
Jul 2, 2025 -
iree-3.6.0rc20250703 iree candidate iree-3.6.0rc20250703
published
Jul 3, 2025
44 Pull requests merged by 21 people
-
[Util] Fix assume.int operand deduplication canonicalizer
#21273 merged
Jul 3, 2025 -
Integrate LLVM to llvm/llvm-project@e3edc1bd
#21272 merged
Jul 3, 2025 -
[ROCM] Ping pong matmul Bf16 matcher
#21267 merged
Jul 2, 2025 -
[integrate|compiler] Drop carried LLVM reverts and use
ub.poisonin some transfer reads#21259 merged
Jul 2, 2025 -
[Codegen] Fix specialize exports never applies check
#21270 merged
Jul 2, 2025 -
Add link to the LLVM Social Bangalore talk
#21265 merged
Jul 2, 2025 -
Integrate LLVM to llvm/llvm-project@5ed852f7
#21263 merged
Jul 2, 2025 -
[DT] fixup(MaterializeEncodingPatterns) remove legacy type conversions
#21262 merged
Jul 1, 2025 -
[mlir][Codegen] Remove workaround for handling consumer fusion along multiple operands.
#21171 merged
Jul 1, 2025 -
[Encoding] Add new identity encoding attribute
#21258 merged
Jul 1, 2025 -
Bump ncipollo/release-action from 1.16.0 to 1.18.0 in the github-actions group
#21254 merged
Jul 1, 2025 -
[CPU] Use scf.forall for TileRootAndFuseProducerConsumer by default.
#21260 merged
Jul 1, 2025 -
[mlir][DispatchCreation] Avoid SSA violation due to consumer fusion while forming dispatches
#21186 merged
Jul 1, 2025 -
[StableHLO] Fix ArrayRef(std::nullopt) deprecation warnings
#21257 8000 merged
Jul 1, 2025 -
[CodeGen] Fix gather fusion on vector distribute path
#21117 merged
Jul 1, 2025 -
[VectorExt] Implement masked vectorization for iree_linalg_ext.gather
#21189 merged
Jul 1, 2025 -
[VectorExt] Implement BufferizationInterface for transfer_gather
#21219 merged
Jul 1, 2025 -
[NFC] Make internal LLVMGPU APIs for vector_distribute available
#21161 merged
Jul 1, 2025 -
[Codegen] Change swizzle hint offset logic to use arith
#21237 merged
Jul 1, 2025 -
[Integrate] Update bufferization related codes for upstream custom types support.
#21250 merged
Jul 1, 2025 -
[CodeGen] Re-enable memref::AssumeAlignmentOp for SPIRV pipelines.
#21133 merged
Jun 30, 2025 -
[Codegen] Add pass to propagate constant offsets towards accesses
#21236 merged
Jun 30, 2025 -
[Integrate] Drop the revert of unknown type conversion in bufferization.
#21243 merged
Jun 30, 2025 -
[ROCMTarget] Make all pingpong arithmetic nsw and nuw
#21248 merged
Jun 30, 2025 -
[ROCm] Fix typo in R9700 SKU definition. NFC.
#21247 merged
Jun 30, 2025 -
[Dispatch Creation] Add concat expand_shape bubbling
#21158 merged
Jun 30, 2025 -
Integrate LLVM to llvm/llvm-project@a99fee69
#21242 merged
Jun 30, 2025 -
Adding AMDGPU semaphore (WIP) and semaphore pool.
#21201 merged
Jun 30, 2025 -
[DispatchCreation] Add pass to hoist scalar ops out of dispatch regions
#21210 merged
Jun 30, 2025 -
[Codegen] Fix TileLargeTensors handling of dynamic reduction dims
#21244 merged
Jun 30, 2025 -
[Codegen][Tuner] expose python binding isa_attention_op
#21216 merged
Jun 30, 2025 -
[Integrate] Drop revert for vectorization API change
#21239 merged
Jun 30, 2025 -
Extending AMDGPU tests, fixing issues, and cleaning up comments.
#21200 merged
Jun 30, 2025 -
[Codegen][GPU] Support fission of loops with multiple transfer_reads/writes
#21213 merged
Jun 30, 2025 -
Adding IREE_ENABLE_RUNTIME_COVERAGE cmake mode.
#21191 merged
Jun 30, 2025 -
[Codegen][GPU] Adding scheduling barrier between compute and write stage in prefetcher
#21151 merged
Jun 30, 2025 -
[LinalgExt] Implement Unit Dim folding for slice dimensions
#21220 merged
Jun 30, 2025 -
Fix missing return in
DeviceOptimalAttr::joinOR#21228 merged
Jun 28, 2025 -
[Codegen] Fix multiple function support in materialize user configs
#21227 merged
Jun 27, 2025 -
Revert "Force install python version 3.13.5 for windows"
#21215 merged
Jun 27, 2025 -
[LinalgExt] add e2e tests for argcompare op
#21217 merged
Jun 27, 2025 -
Integrate LLVM @ c73e5e3e209c
#21224 merged
Jun 27, 2025 -
[NFC] remove redundant checks in the TilingInterface
#21225 merged
Jun 27, 2025 -
[Codegen][GPU] Move LLVMGPUPrefetching pass to be invoked from only amdgpu backend
#21190 merged
Jun 27, 2025
20 Pull requests opened by 11 people
-
[CodeGen] Add a pass that patches func ops for debugging purpose.
#21229 opened
Jun 27, 2025 -
Bump torch-mlir to 8cc313fd0d8882a44b34c5bcf3d32a0d0089161b
#21233 opened
Jun 27, 2025 -
[CPU] Use rootOp in distribution for mmt4d pipeline.
#21234 opened
Jun 27, 2025 -
[Codegen] Fix lhs/rhs batch offsets size in vector contract distribution
#21238 opened
Jun 30, 2025 -
[Codegen] Bubble up/down reshape operations before blocking dynamic dimensions
#21241 opened
Jun 30, 2025 -
[Codegen] Support inner_tiled and load_from_buffer in ConvertAccGEMMToGEMMPass
#21245 opened
Jun 30, 2025 -
[Codegen] Generalize ukernel strided_outer_dims and fix GPU ukernel bug
#21249 opened
Jun 30, 2025 -
Implementing AMDGPU logical/physical devices and skeleton queues.
#21251 opened
Jun 30, 2025 -
[Im2col] Fuse pad op into im2col decomposition loops
#21255 opened
Jul 1, 2025 -
[Codegen][GPU] Redirect transfer read lowering to masked load lowering
#21269 opened
Jul 2, 2025 -
[LinalgExt] Improve scatter unit dim folding
#21271 opened
Jul 2, 2025 -
[DispatchCreation] Fold collapse(expand) unit dims
#21274 opened
Jul 2, 2025 -
[DT] Improve encoding hoisting pass.
#21275 opened
Jul 3, 2025 -
[DT] add control function to FoldIntoPackUnpackPatterns
#21276 opened
Jul 3, 2025 -
[TensorExt] Mirror a tensor_ext version of flow.tensor.bitcast
#21277 opened
Jul 3, 2025 -
Move all producing uses of flow.bitcast to tensor_ext
#21279 opened
Jul 3, 2025 -
[DispatchCreation] Add a pass to split long running reduction loops.
#21280 opened
Jul 3, 2025 -
[Codegen] Resolve scf.forall operations created during split-reduction.
#21283 opened
Jul 4, 2025
12 Issues closed by 11 people
-
iree-compile produces invalid IR form valid input (layernorm-ish)
#21268 closed
Jul 2, 2025 -
Tracking issue for SSA use-def violation issue with tile and fuse consumer
#21091 closed
Jul 1, 2025 -
Concatting tensors causes resulting values to be different
#21087 closed
Jul 1, 2025 -
[Encoding] Missing identity encoding attribute
#21183 closed
Jul 1, 2025 -
[Dispatch Creation] sdxl_unet_fp16 without preprocessing
#20191 closed
Jul 1, 2025 -
[codegen] Wan2.1 transformer domination issue with aggressive fusion
#21176 closed
Jul 1, 2025 -
Shared memory failure on softmax
#21107 closed
Jul 1, 2025 -
[Integrate] Fail to hoisting vectors when memref.assume_alignment is present
#20912 closed
Jun 30, 2025 -
[DispatchCreation] Support reshape propagation for tensor.concat
#21154 closed
Jun 30, 2025 -
Dimension mismatch after the bubble-up-expand-shapes pass
#21033 closed
Jun 30, 2025 -
[Stream] Crash in `AffinityAttr::joinOR`
#21203 closed
Jun 28, 2025
11 Issues opened by 8 people
-
Assertion `problem.aType == problem.bType && "expected the same aType and bType."' failed
#21278 opened
Jul 3, 2025 -
Maybe a bug in KernelConfig.cpp
#21266 opened
Jul 2, 2025 -
Matmul performance drop when using expanded shape
#21256 opened
Jul 1, 2025 -
[GPU][DT] Propagate unset_encoding through elementwise generic
#21253 opened
Jul 1, 2025 -
[GPU][DT] Stack frame size exceeded error
#21252 opened
Jul 1, 2025 -
[DT] Be able to selectively enable data-tiling with hints
#21246 opened
Jun 30, 2025 -
[CPU] Fail to vectorize unpack ops in data-tiling fusion
#21235 opened
Jun 27, 2025 -
MXFP4 Llama3.1 405b Bring Up
#21232 opened
Jun 27, 2025 -
Compiler crash when memoizing a function
#21231 opened
Jun 27, 2025 -
Compiler crash when hoisting Flow ops
#21230 opened
Jun 27, 2025 -
[Attention] Paged Attention Codegen
#21222 opened
Jun 27, 2025
19 Unresolved conversations
Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.
-
[CODEGEN] Updated map_scatter insertion in CombineLayoutTransform Pass
#21206 commented on
Jul 1, 2025 • 12 new comments -
[DispatchCreation] Form splitk dispatches
#21192 commented on
Jun 30, 2025 • 6 new comments -
[Codegen] Support dynamic dimensions in collapse_shape into interface store folding
#21126 commented on
Jul 2, 2025 • 4 new comments -
Pad OnlineAttention
#21152 commented on
Jul 3, 2025 • 1 new comment -
[HAL] Add hal.allocator.resolve_memory_properties
#21115 commented on
Jul 1, 2025 • 1 new comment -
WIP - Switching to data-tiling fusion.
#21181 commented on
Jul 3, 2025 • 0 new comments -
[Codegen] Account for pre-distributed computation in workgroup tiling
#21137 commented on
Jul 1, 2025 • 0 new comments -
[compiler] Removing copies of duplicate globals
#21130 commented on
Jul 1, 2025 • 0 new comments -
[Stream] New ElideAsyncTransfersPass
#21029 commented on
Jul 1, 2025 • 0 new comments -
[GPU] Vector distribute support more dynamic cases
#20959 commented on
Jul 1, 2025 • 0 new comments -
[CPU] linalg.pack op is not fused in forall distribution
#20723 commented on
Jul 3, 2025 • 0 new comments -
Llama_405b_tp8 OOM w/ Long Input Prompt
#19832 commented on
Jul 2, 2025 • 0 new comments -
Implement initial heterogeneous support MVP for CPU (+ maybe something else).
#20851 commented on
Jul 2, 2025 • 0 new comments -
iree-compile LLVM ERROR: operation destroyed but still has uses on ScheduleExecutionPass
#20354 commented on
Jul 1, 2025 • 0 new comments -
Fix Global Opt for 1x1 Filter Convs
#19270 commented on
Jun 30, 2025 • 0 new comments -
[CPU] CPU backend produces unfused IR in mmt4d->unpack->elem dispatch
#20786 commented on
Jun 27, 2025 • 0 new comments -
Compiler crash during SPIRVInitialVectorLowering
#20579 commented on
Jun 27, 2025 • 0 new comments -
[Mistral] Performance degradation with VMFB containing prefill functions of multiple batch sizes
#20836 commented on
Jun 27, 2025 • 0 new comments -
[DispatchCreation] Failed to hoist out of dispatch in DT-fusion
#21100 commented on
Jun 27, 2025 • 0 new comments